MHB Looking for a parameter that expresses quality of spatial distribution

AI Thread Summary
A mathematical parameter is sought to quantify the quality of spatial distribution of points in a three-dimensional space, specifically to indicate how evenly they are spread. The desired parameter should yield high values for grid-like distributions and low values for clustered arrangements. The discussion notes that the problem is complex, as there are various distribution types beyond just grid and clumped, such as random but uniform distributions. Additional context about the application or specific requirements could enhance the search for a suitable solution. The conversation emphasizes the need for a clear metric to assess point distribution effectively.
corsica
Messages
1
Reaction score
0
My situation is as follows: I have a set of points in a bounded three-dimensional space. Simply put, each point has an X, Y and Z coordinate.

I'm looking for a mathematical parameter that (globally) expresses how well the points are distributed over the space. The parameters should take the X, Y and Z coordinates as inputs. The parameter should return a high value when the points are equally spaced in a grid-like fashion. On the other hand the parameter should return a low value when many of the points are congregated into lumps.

I'm been through all my mathematics books, but I couldn't find a parameter that does this sort of thing. Does anyone here have a solution?
 
Physics news on Phys.org
corsica said:
My situation is as follows: I have a set of points in a bounded three-dimensional space. Simply put, each point has an X, Y and Z coordinate.

I'm looking for a mathematical parameter that (globally) expresses how well the points are distributed over the space. The parameters should take the X, Y and Z coordinates as inputs. The parameter should return a high value when the points are equally spaced in a grid-like fashion. On the other hand the parameter should return a low value when many of the points are congregated into lumps.

I'm been through all my mathematics books, but I couldn't find a parameter that does this sort of thing. Does anyone here have a solution?

The problem is slightly ill posed since there are alternatives to a grid like arrangement and clumped, how about random but with a uniform distribution.

It might help if you could provide some context so we can get a better handle on what you want to measure.

CB
 
Back
Top